gms | German Medical Science

Background and aims: Anticoagulation after acute ischaemic stroke (IS) or transient ischaemic attack (TIA) in patients with atrial fibrillation (AF) increases the risk for bleeding complications, especially in the early setting. The Prospective Record of the Use of Dabigatran in Patients with Acute Stroke or TIA (PRODAST) study was initiated to provide real-world data on antithrombotic treatment and clinical outcomes, with a focus on major bleeding complications. We use these data to validate and improve the simple and established HAS-BLED-Score of major bleeding risk [1].

Methods: PRODAST is a prospective, multi-centre, observational post-authorization safety study that recruited over 10,000 patients with recent (≤1 week) IS or TIA and non-valvular AF at 86 German stroke units, providing real world data on antithrombotic treatment and clinical outcomes. All patients, except those treated with non-vitamin-K–dependent oral anticoagulants (NOAC) other than dabigatran underwent central follow-up after three months. Details of medical history (history of major hemorrhage, hypertension, abnormal liver or renal function, risk of falls, diabetes mellitus, dyslipidemia, malignant diseases, anti-thrombotic treatment and coagulation disorders) and other risk factors such as alcohol consumption, smoking (past or present), adiposity, age and sex were collected at study inclusion, together with size and severity of the index stroke. Cox-regression was applied to model early bleeding risk, using bootstrap resampling for internal validation. Validation was focused on discrimination (Harrell’s C).

Results: From 9655 patients with study data, 289 patients experienced a major bleeding event ≤7 days after the index event, 2412 patients were censored due to death (118) or other major adverse events (215), or due to discharge (1893) or loss to follow-up (186) (event rate 48.0/10,000 patient days). The HAS-BLED score at admission (including 1 point for index stroke) showed a mean of 3.3 (std=0.8, max=5) among the early bleeding cases), which was only slightly higher than the mean of 3.2 (std 0.8, max=6) among the other patients (<1% missing values). Accordingly, Harrell’s C=53.8% gives a very low discriminative potential of the HAS-BLED score in this sample. Remodeling risk of early major bleeding with the single risk factors, including age and drinks/week continuously (log-linear), improved discrimination to C=60.1%. Addition of the risk factors smoking, body-mass index, sex, size and severity of index stroke, and history of risk of falls, diabetes mellitus, dyslipidemia, malignant disease and low platelet count further improved discrimination to C=71.7%.

Conclusions: The HAS-BLED score is simple to use and may have greater merit in a more diverse and healthier population. However, we see a large potential to improve discrimination between persons at high and low risk of major bleeding among patients in the early phase after acute IS or TIA, using information on the patients’ clinical history and severity of the recent event.

Funding: The investigator initiated PRODAST-study was funded by an unrestricted grant from Boehringer Ingelheim.

The authors declare that a positive ethics committee vote has been obtained.